DESCRIPTION (Applicant's Abstract): The last decade has seen great progress in understanding which people with mental illness are likely to be involved in violent incidents. Although this is an impressive achievement, it will do little to help manage the violence because we can no longer hospitalize potentially violent patients for long periods of time. If we are to help minimize the violent in the lives of the mentally ill in the community, we must develop an understanding of the origins and patterns of behavior in the violent incidents themselves. This exploratory research will study characteristics of violent incidents rather than violent people. The investigators will interview both the patient and the co-participant in 100 violent incidents in order to: 1) assess the role of the mental illness in these incidents; 2) analyze the sequences of emotions and conflict tactics used; 3) compare these patterns to the sequences of emotions and conflict tactics in non-violent conflicts between the same two people; 4) identify the types of justifications and excuses for the use of violence made by each party; and 5) describe the perceived strains in the relationships between the parties to the violent incidents. Each of these aims will directly facilitate a different approach to treating or managing violence involving individuals with mental illness. The first will inform us about the likely consequences of successfully treating the mental illness. The second and third aims are intended to facilitate the design of behavioral interventions. The fourth aim addresses the possibility of affecting violence by direct confrontation of the violent individual's excuses as is often done in groups for spouse abusers. The final aim is intended to facilitate group and family therapy approaches to violence management.